Vehicular Signal Acquisition and Processing

Team members:

James McCrae - Laptop interfacing and software subsystem
Bryan Stuurman - Mechanical and electrical hardware

Supervisor:

Arnold Rosenbloom

Objective:

The development of a computer system which is physically connected to the sensors of the vehicle's motor to acquire signals. Also, the development of a software solution to acquiring and processing the data from the connection. Finally, an open exploration into the relevance and potential for this system - and Computer Science in automotive technology.

Current Status:

• Week 1
  • Signal conditioner module planning and implementation for the Tachometer can be found here.
• Week 2
  • Initial skeleton code for software subsystem can be found here and the package can be downloaded here.
• Week 3
  • Preliminary simulation for software subsystem implemented. Code found here, package can be downloaded here.
• Week 4
  • Revised signal conditioner module planning for the Tachometer here.
  • Signal conditioner module planning for the Speed Sensor here.
  • Implementation of both modules found here.
• Week 5
  • Implementation documentation for the 4-Bit Counters and Signal Bus can be found here.
• Week 6
  • Revised software implementation here, compiled package here.
  • A high-level discussion of the approach to creating the software found here.
  • Discussion on the class architecture of the software subsystem here.
• Week 7
  • Software implementation including speedometer functionality here, compiled package here.
• Week 8
  • A discussion on determining an approximation of the vehicle's power output can be found here.
• Week 9
  • Software implementation which includes graphical Power Output analysis can be found here, compiled package here.
• Week 10
  • A revised software implementation with graphical outputs for velocity and acceleration over time can be found here, compiled package here.
  • A discussion on an alternate method for estimating the power output can be found here.
• Week 11
  • Various screen captures of the software found here.
  • Software package with averaging smoothing function found here.
• Week 12
  • Software package with dynamometer, configuration file implementation and dynamometer file import/export functionality found here.
• Week 13
  • This week's software package contains an implementation that approximates frictional forces acting on the vehicle within the dynamometer feature to produce a more accurate horsepower curve. The package can be found here.

Schedule of Deliverables:

June 4/04	Project Commencement
June 15/04	Custom Signal Conditioner Modules constructed and installed in vehicle
June 18/04	Preliminary software subsystem design complete (basic interface, simulating sensor inputs, and preliminary signal processing for tach signal and vehicle speed displays)
June 25/04	Signal bus constructed and installed in vehicle
June 30/04	Laptop computer installed in vehicle and parallel port connected to signal bus
June 30/04	Hardware communication modules for software subsystem prepared for on-board testing
July 7/04	Worked out particulars/difficulties with intercommunication of hardware/software subsystems
July 21/04	Finalized signal processing functionalities of software subsystem and error detection features

Finally, dependent on time constraints and project requirements, acquire additional signals, and add or extend features.

Technical Overview:

The selected vehicle is a 1983 Toyota Celica, equipped with a 2.4L 4 cylinder reciprocating gasoline combustion engine. The engine is controlled via an analog computer, which monitors up to 7 sensors to determine the proper air-fuel ratio for the specific driving circumstance. Spark timing is separately controlled via a mechano-vaccum advancing system completely isolated from the Engine Control Unit (ECU).

The computer selected is a Commodore laptop, model ANC-4S25. It has a 486 33MHz DX processor (overclocked to 40MHz), Chips 65520 VGA video controller, 640KB base memory and 7424KB extended memory (8MB RAM total), Toshiba HDD2339 250MB Hard Drive, 1.44MB 3.5" disk drive, parallel port (hardware address 0x378) and serial port.

The choice of the vehicle and computer system were severely limited by availability of test equipment, however, these platforms offer an exciting challenge in interfacing, and allow exploration of Computer Science solutions to signal acquisition and processing.

Project Plan:

Stage 1

In stage one, our goal is to install the entire laptop computer into the car, and interface its parallel port to the vehicle's sensors. The four TTL input lines on the parallel port will be treated as a bus. 4-bit counters driven by conditioned signals from important sensors can be selectively attached to the bus and polled for the sensor's "count". Such a system allows for up to 6 input sensors with 4-bit resolution. The 12 TTL outputs on the parallel port will serve as data select/reset pairs for the six input devices.

The parallel port is a TTL I/O port which is completely software driven. This allows for expected polling rates of up to a 1% maximum of the CPU clock speed (assuming a C implementation, in a monotasking enviroment). A polling rate of at least 400kHz is then expected of the test computer, which is more than adequate considering the reduction of overhead given by the 4-bit counters. The net effect is a 400*16 kHz, or 6.4MHz sample rate. Clearly, such a sample rate is not required, but the speed of data acquisition leaves CPU overhead and thus potential for computation in real time. This hardware configuration ensures that signals can be acquired faster than they are given.

Fundamental sensors selected for observation will be Tachometer and Speed. The Tachometer signal is a "messy" digital signal with about the same levels as 12v CMOS signalling. This signal will need to be inverted, dropped to a TTL level and conditioned with a small high pass filter to eliminate transients. The Speed signal is somewhat sinusoidal at variable voltage and frequency. A conditioner module will be built to detect the zero-crossings of the sinusoid and convert them to TTL logic pulses (since there are two pulses for each wave, the acquired value will have to be divided by two - to be addressed by the software subsystem).

The following diagram presents the modular components of the electrical hardware system:

Stage 2

Software for the assembled hardware platform from stage one will consist of MS-DOS as the operating system, with an implementation built in either DJGPP (MS-DOS port of GNU C/C++ compiler) or Borland Turbo C/C++. Initial code will be designed simply to test I/O routines and ensure the data from the vehicle sensors is correctly acquired. This can be determined by ensuring the data displayed by the computer is coincident with the readings presented on the dashboard of the vehicle.

Upon diagnosing and addressing all hardware issues, the software subsystem will have successfully acquired and displayed signal information.

Stage 3

This stage provides an open exploration into the possibilities available with the system implementation. Research and analysis will be conducted, and solutions will be proposed to the following issues:

• Discretization and granularity of the input
• Performance of the software subsystem concurrently acquiring and processing signals
• Data that can be processed, derived and presented from the input signals
• Development of a simplistic UI targeted toward a vehicle operator

References:

"Interfacing the IBM PC Parallel Printer Port"
http://www.lvr.com/files/ibmlpt.txt

"Programming the parallel port"
http://www.gmonline.demon.co.uk/cscene/CS4/CS4-02.html

"How Car Computers Work"
http://auto.howstuffworks.com/car-computer.htm

"1983 Celica Maintenance Manual"
Toyota Motor Corp.

"DJGPP"
http://www.delorie.com/djgpp/